Modelling of Dry-Low Emission Gas Turbine Fuel System using First Principle Data-Driven Method

Achieving reliable power generation from Dry Low Emission gas turbines together with low CO2 and NOx discharge is a great challenge, as the rigorous control strategy is susceptible to frequent trips. Therefore, it is crucial to establish a dynamic model of the turbine (such as the one commonly attributed to Rowen) to ascertain the stability of the system. However, the major distinctive fuel system design in the DLE gas turbine is not constructed in the well-established model. With this issue in mind, this paper proposes a modelling approach to the DLE gas turbine fuel system which consists of integrating the main and pilot gas fuel valve into Rowen’s model, using the First Principle Data-Driven (FPDD) method. First, the structure of the fuel system is determined and generated in system identification. Subsequently, the validated valve models are integrated into Rowen’s model as the actual setup of the DLE gas turbine system. Ultimately, the core of this modelling approach is fuel system integration based on the FPDD method to accurately represent the actual signals of the pilot and main gas fuel valves, gas fuel flow and average turbine temperature. Then, the actual signals are used to validate the whole structure of the model using MAE and RMSE analysis. The results demonstrate the high accuracy of the DLE gas turbine model representation for future utilization in fault identification and prediction study.

---

Opracowanie rekordu ze środków MNiSW, umowa Nr 461252 w ramach programu "Społeczna odpowiedzialność nauki" - moduł: Popularyzacja nauki i promocja sportu (2020).